gsadmin/spinifex
1
0
Fork 0
mirror of https://github.com/mulgadc/spinifex.git synced 2026-10-15 09:01:31 +00:00
Code Issues Packages Projects Releases Wiki Activity

refactor: establish node VM inventory contract

Browse Source
This commit is contained in:
joshsiv-mulga
2026-09-30 16:27:39 +10:00
parent 78fd017e03
commit 4b35a2a33d
16 changed files with 164 additions and 82 deletions
Download Patch File Download Diff File Expand all files Collapse all files
cmd/spinifex/cmd/cluster.go
+4 -4
View File
@@ -10,10 +10,10 @@ import (
"strings"
"time"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/spinifex/daemon"
"github.com/mulgadc/spinifex/spinifex/network/external/dhcp"
"github.com/mulgadc/spinifex/spinifex/foundation/telemetry"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/mulgadc/spinifex/spinifex/network/external/dhcp"
"github.com/nats-io/nats.go"
"github.com/spf13/cobra"
)
@@ -376,7 +376,7 @@ func reportGuestsLeftRunning(severe bool, msg string) {
defer nc.Close()
node := cfg.Node
responses, err := collectResponses(nc, "spinifex.node.vms", guestEnumerationTimeout)
responses, err := collectResponses(nc, clusterv1.NodeVMsSubject, guestEnumerationTimeout)
if err != nil {
slog.Warn("could not enumerate local guests", "node", node, "error", err)
return
@@ -384,7 +384,7 @@ func reportGuestsLeftRunning(severe bool, msg string) {
var running []string
for _, data := range responses {
var resp types.NodeVMsResponse
var resp clusterv1.NodeVMsResponse
if err := json.Unmarshal(data, &resp); err != nil {
continue
}
cmd/spinifex/cmd/cluster_test.go
+11 -11
View File
@@ -8,11 +8,11 @@ import (
"testing"
"time"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/internal/testkit"
"github.com/mulgadc/spinifex/spinifex/config"
"github.com/mulgadc/spinifex/spinifex/daemon"
"github.com/mulgadc/spinifex/spinifex/network/external/dhcp"
"github.com/mulgadc/spinifex/internal/testkit"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/nats-io/nats.go"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
@@ -287,10 +287,10 @@ func TestWarnIfGuestsLeftRunningNamesRunningGuests(t *testing.T) {
_, nc, _ := testutil.StartTestJetStream(t)
stubConnect(t, &config.ClusterConfig{Node: "node1"}, nc, nil)
sub, err := nc.Subscribe("spinifex.node.vms", func(msg *nats.Msg) {
resp := types.NodeVMsResponse{
sub, err := nc.Subscribe(clusterv1.NodeVMsSubject, func(msg *nats.Msg) {
resp := clusterv1.NodeVMsResponse{
Node: "node1",
VMs: []types.VMInfo{
VMs: []clusterv1.VMInfo{
{InstanceID: "i-running", Status: vmStatusRunning},
{InstanceID: "i-stopped", Status: "stopped"},
},
@@ -318,10 +318,10 @@ func TestWarnIfGuestsLeftRunningNoGuestsIsSilent(t *testing.T) {
_, nc, _ := testutil.StartTestJetStream(t)
stubConnect(t, &config.ClusterConfig{Node: "node1"}, nc, nil)
sub, err := nc.Subscribe("spinifex.node.vms", func(msg *nats.Msg) {
resp := types.NodeVMsResponse{
sub, err := nc.Subscribe(clusterv1.NodeVMsSubject, func(msg *nats.Msg) {
resp := clusterv1.NodeVMsResponse{
Node: "node1",
VMs: []types.VMInfo{{InstanceID: "i-stopped", Status: "stopped"}},
VMs: []clusterv1.VMInfo{{InstanceID: "i-stopped", Status: "stopped"}},
}
body, _ := json.Marshal(resp)
_ = msg.Respond(body)
@@ -516,10 +516,10 @@ func TestReportGuestsLeftRunningSevereLogsErrorAndNamesGuests(t *testing.T) {
_, nc, _ := testutil.StartTestJetStream(t)
stubConnect(t, &config.ClusterConfig{Node: "node1"}, nc, nil)
sub, err := nc.Subscribe("spinifex.node.vms", func(msg *nats.Msg) {
resp := types.NodeVMsResponse{
sub, err := nc.Subscribe(clusterv1.NodeVMsSubject, func(msg *nats.Msg) {
resp := clusterv1.NodeVMsResponse{
Node: "node1",
VMs: []types.VMInfo{
VMs: []clusterv1.VMInfo{
{InstanceID: "i-abandoned", Status: vmStatusRunning},
{InstanceID: "i-stopped", Status: "stopped"},
},
cmd/spinifex/cmd/get.go
+4 -3
View File
@@ -13,6 +13,7 @@ import (
"strings"
"time"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/spinifex/config"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/mulgadc/spinifex/spinifex/utils"
@@ -241,14 +242,14 @@ func runGetVMs(cmd *cobra.Command, args []string) {
defer nc.Close()
timeout, _ := cmd.Flags().GetDuration("timeout")
responses, err := collectResponses(nc, "spinifex.node.vms", timeout)
responses, err := collectResponses(nc, clusterv1.NodeVMsSubject, timeout)
if err != nil {
fmt.Fprintf(os.Stderr, "Error: %v\n", err)
os.Exit(1)
}
type vmRow struct {
types.VMInfo
clusterv1.VMInfo
Node string
Host string
@@ -256,7 +257,7 @@ func runGetVMs(cmd *cobra.Command, args []string) {
var allVMs []vmRow
for _, data := range responses {
var resp types.NodeVMsResponse
var resp clusterv1.NodeVMsResponse
if err := json.Unmarshal(data, &resp); err != nil {
continue
}
contracts/cluster/v1/README.md
+7
View File
@@ -13,3 +13,10 @@ when a wire change is required.
`NodeHealthResponse` is also the daemon HTTPS `/health` representation. Do not
change their subject or JSON shape incompatibly in this package; add a new
version when a wire change is required.
`NodeVMsSubject` retains the deployed `spinifex.node.vms` fan-out route.
`NodeVMsResponse` is the VM inventory reply used by the SPX operator CLI and
by EKS host lookup. It is distinct from node status and capacity reporting:
do not add scheduling or physical-GPU inventory merely because those callers
also inspect nodes. Do not change its subject or JSON shape incompatibly; add
a new version when a wire change is required.
contracts/cluster/v1/node_vms.go
+44
View File
@@ -0,0 +1,44 @@
package clusterv1
const (
// NodeVMsSubject fans out a request for each responding daemon's VM
// inventory. Its reply payload is NodeVMsResponse.
NodeVMsSubject = "spinifex.node.vms"
)
// VMGPUInfo describes one GPU attached to a VM in a NodeVMsResponse.
type VMGPUInfo struct {
Model string `json:"model"`
VRAMMiB int64 `json:"vram_mib"`
PCIAddress string `json:"pci_address,omitempty"` // whole-GPU passthrough only
Profile string `json:"profile,omitempty"` // MIG profile name; empty for whole-GPU
MdevPath string `json:"mdev_path,omitempty"` // MIG only
}
// VMInfo describes one VM in a node's inventory response.
type VMInfo struct {
InstanceID string `json:"instance_id"`
Status string `json:"status"`
InstanceType string `json:"instance_type"`
VCPU int `json:"vcpu"`
MemoryGB float64 `json:"memory_gb"`
LaunchTime int64 `json:"launch_time"`
// ManagedBy is the Spinifex platform component that owns this VM
// (e.g. "elbv2"). Empty for customer VMs. The UI uses this to filter
// system-managed resources out of customer-facing listings.
ManagedBy string `json:"managed_by,omitempty"`
GPUs []VMGPUInfo `json:"gpus,omitempty"`
// Health is a display label for instance health: "ok", "impaired",
// "recovering", or "-" for non-running VMs. CrashCount is the lifetime
// crash tally within the current restart window.
Health string `json:"health,omitempty"`
CrashCount int `json:"crash_count,omitempty"`
}
// NodeVMsResponse is one daemon's reply on NodeVMsSubject. The subject fans
// out, so callers collect one response per responding node.
type NodeVMsResponse struct {
Node string `json:"node"`
Host string `json:"host"`
VMs []VMInfo `json:"vms"`
}
contracts/cluster/v1/node_vms_test.go
+60
View File
@@ -0,0 +1,60 @@
package clusterv1
import (
"encoding/json"
"testing"
"github.com/stretchr/testify/require"
)
func TestNodeVMsResponseJSONContract(t *testing.T) {
data, err := json.Marshal(NodeVMsResponse{
Node: "node-a",
Host: "10.0.0.1",
VMs: []VMInfo{{
InstanceID: "i-abc123",
Status: "running",
InstanceType: "t3.small",
VCPU: 2,
MemoryGB: 2.0,
LaunchTime: 1736935200,
ManagedBy: "elbv2",
GPUs: []VMGPUInfo{{
Model: "NVIDIA A10",
VRAMMiB: 23028,
PCIAddress: "0000:01:00.0",
Profile: "MIG 1g.6gb",
MdevPath: "/sys/bus/mdev/devices/example",
}},
Health: "ok",
CrashCount: 2,
}},
})
require.NoError(t, err)
require.JSONEq(t, `{
"node":"node-a",
"host":"10.0.0.1",
"vms":[{
"instance_id":"i-abc123",
"status":"running",
"instance_type":"t3.small",
"vcpu":2,
"memory_gb":2,
"launch_time":1736935200,
"managed_by":"elbv2",
"gpus":[{
"model":"NVIDIA A10",
"vram_mib":23028,
"pci_address":"0000:01:00.0",
"profile":"MIG 1g.6gb",
"mdev_path":"/sys/bus/mdev/devices/example"
}],
"health":"ok",
"crash_count":2
}]
}`, string(data))
}
func TestNodeVMsSubject(t *testing.T) {
require.Equal(t, "spinifex.node.vms", NodeVMsSubject)
}
docs/PACKAGE_BOUNDARY_MIGRATION.md
+1
View File
@@ -186,6 +186,7 @@ while the branch is in flight; they do not describe the umbrella repository's
| `c239fd897` | `spinifex/types/eni.go` | `spinifex/runtime/compute/vm/eni_requests.go` | PCIe hot-plug slot allocation is mutex-bearing, per-VM runtime state rather than a shared type or wire contract; `record_test.go` preserves its restart semantics. |
| `6d6275e78` | `spinifex/types/cluster.go` (`NodeHealthResponse`) | `contracts/cluster/v1/health.go` | Node health is a versioned cluster control-plane representation shared by the per-node NATS route and daemon HTTPS `/health`; `health_test.go` pins both wire forms. |
| `53a614717` | `spinifex/types/cluster.go` (`SharedClusterData`) | `spinifex/daemon/config_hash.go` | Shared-config hash input is daemon-local implementation data; focused config-hash tests preserve deterministic, shared-only semantics. |
| `78fd017e0` | `spinifex/types/telemetry.go` (`GuestTelemetryMeta`) | `contracts/telemetry/v1/guest_meta.go` | QMP collector sidecar JSON is a versioned file contract between VM runtime and telemetry role; `guest_meta_test.go` pins its filename and payload. |
The EC2 contract row is intentionally different from the directory moves: it creates
a compatibility boundary. `ec2.cmd.*` retains its deployed one-token NATS
docs/service-interfaces.yaml
+1
View File
@@ -92,6 +92,7 @@ services:
# rarely the node that served the request (fan-out, no queue group).
- "ec2.eni.public-ip-changed"
- "spinifex.node.status"
# contracts/cluster/v1 owns the VM inventory payload and subject.
- "spinifex.node.vms"
- "spinifex.nodes.discover"
- "spinifex.cluster.shutdown.progress"
spinifex/daemon/daemon.go
+1 -1
View File
@@ -1021,7 +1021,7 @@ func (d *Daemon) subscribeAll() error {
{clusterv1.NodeHealthSubject(d.node), d.handleHealthCheck, ""},
{clusterv1.NodesDiscoverSubject, d.handleNodeDiscover, ""},
{"spinifex.node.status", d.handleNodeStatus, ""},
{"spinifex.node.vms", d.handleNodeVMs, ""},
{clusterv1.NodeVMsSubject, d.handleNodeVMs, ""},
{operatorv1.StorageConfigSubject, d.handleStorageConfig, ""},
{"spinifex.image.promote", d.handleSpinifexPromoteImage, "spinifex-workers"},
// Account creation → create default VPC for new account
spinifex/daemon/daemon_gpu_handlers_test.go
+4 -3
View File
@@ -7,9 +7,10 @@ import (
"testing"
"time"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/spinifex/runtime/compute/gpu"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/mulgadc/spinifex/spinifex/runtime/compute/vm"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
@@ -31,7 +32,7 @@ func gpuStatusRequest(t *testing.T, daemon *Daemon, subject string) types.NodeSt
}
// gpuVMsRequest subscribes and requests handleNodeVMs on a fresh subject.
func gpuVMsRequest(t *testing.T, daemon *Daemon, subject string) types.NodeVMsResponse {
func gpuVMsRequest(t *testing.T, daemon *Daemon, subject string) clusterv1.NodeVMsResponse {
t.Helper()
sub, err := daemon.natsConn.Subscribe(subject, asMsgHandler(daemon.handleNodeVMs))
require.NoError(t, err)
@@ -40,7 +41,7 @@ func gpuVMsRequest(t *testing.T, daemon *Daemon, subject string) types.NodeVMsRe
reply, err := daemon.natsConn.Request(subject, nil, 5*time.Second)
require.NoError(t, err)
var resp types.NodeVMsResponse
var resp clusterv1.NodeVMsResponse
require.NoError(t, json.Unmarshal(reply.Data, &resp))
return resp
}
spinifex/daemon/daemon_handlers.go
+8 -8
View File
@@ -578,10 +578,10 @@ func buildPoolLookup(mgr *gpu.Manager) (byMdev, byPCI map[string]gpu.PoolEntry)
// resolveVMGPU maps a single GPUAttachment to a VMGPUInfo using the pool
// lookup tables built by buildPoolLookup. Returns nil if the attachment cannot
// be matched (e.g. daemon restart before pool is fully restored).
func resolveVMGPU(att gpu.GPUAttachment, byMdev, byPCI map[string]gpu.PoolEntry) *types.VMGPUInfo {
func resolveVMGPU(att gpu.GPUAttachment, byMdev, byPCI map[string]gpu.PoolEntry) *clusterv1.VMGPUInfo {
if att.MdevPath != "" {
if e, ok := byMdev[att.MdevPath]; ok && e.MIGInstance != nil {
return &types.VMGPUInfo{
return &clusterv1.VMGPUInfo{
Model: e.Device.Model,
VRAMMiB: e.MIGInstance.Profile.MemoryMiB,
Profile: e.MIGInstance.Profile.Name,
@@ -592,7 +592,7 @@ func resolveVMGPU(att gpu.GPUAttachment, byMdev, byPCI map[string]gpu.PoolEntry)
}
if att.PCIAddress != "" {
if e, ok := byPCI[att.PCIAddress]; ok {
return &types.VMGPUInfo{
return &clusterv1.VMGPUInfo{
Model: e.Device.Model,
VRAMMiB: e.Device.MemoryMiB,
PCIAddress: att.PCIAddress,
@@ -605,8 +605,8 @@ func resolveVMGPU(att gpu.GPUAttachment, byMdev, byPCI map[string]gpu.PoolEntry)
// resolveVMGPUs resolves every attachment it can, reporting a short list rather
// than passing it off as complete. One unresolved attachment used to read as no
// GPU at all; eight make a shortened list a plausible wrong answer.
func resolveVMGPUs(instanceID string, attachments []gpu.GPUAttachment, byMdev, byPCI map[string]gpu.PoolEntry) []types.VMGPUInfo {
gpus := make([]types.VMGPUInfo, 0, len(attachments))
func resolveVMGPUs(instanceID string, attachments []gpu.GPUAttachment, byMdev, byPCI map[string]gpu.PoolEntry) []clusterv1.VMGPUInfo {
gpus := make([]clusterv1.VMGPUInfo, 0, len(attachments))
for _, attachment := range attachments {
if info := resolveVMGPU(attachment, byMdev, byPCI); info != nil {
gpus = append(gpus, *info)
@@ -627,9 +627,9 @@ func resolveVMGPUs(instanceID string, attachments []gpu.GPUAttachment, byMdev, b
func (d *Daemon) handleNodeVMs(msg *nats.Msg) string {
poolByMdev, poolByPCI := buildPoolLookup(d.gpuManager)
vms := make([]types.VMInfo, 0, d.vmMgr.Count())
vms := make([]clusterv1.VMInfo, 0, d.vmMgr.Count())
d.vmMgr.ForEach(func(v *vm.VM) {
info := types.VMInfo{
info := clusterv1.VMInfo{
InstanceID: v.ID,
Status: string(v.Status),
InstanceType: v.InstanceType,
@@ -648,7 +648,7 @@ func (d *Daemon) handleNodeVMs(msg *nats.Msg) string {
vms = append(vms, info)
})
resp := types.NodeVMsResponse{
resp := clusterv1.NodeVMsResponse{
Node: d.node,
Host: d.daemonIP(),
VMs: vms,
spinifex/daemon/daemon_handlers_test.go
+4 -4
View File
@@ -2239,7 +2239,7 @@ func TestHandleNodeVMs(t *testing.T) {
reply, err := daemon.natsConn.Request("spinifex.node.vms.test", nil, 5*time.Second)
require.NoError(t, err)
var resp types.NodeVMsResponse
var resp clusterv1.NodeVMsResponse
err = json.Unmarshal(reply.Data, &resp)
require.NoError(t, err)
@@ -2248,7 +2248,7 @@ func TestHandleNodeVMs(t *testing.T) {
assert.Len(t, resp.VMs, 2)
// Build a lookup by instance ID
vmsByID := make(map[string]types.VMInfo)
vmsByID := make(map[string]clusterv1.VMInfo)
for _, v := range resp.VMs {
vmsByID[v.InstanceID] = v
}
@@ -2276,7 +2276,7 @@ func TestHandleNodeVMs_Empty(t *testing.T) {
reply, err := daemon.natsConn.Request("spinifex.node.vms.empty", nil, 5*time.Second)
require.NoError(t, err)
var resp types.NodeVMsResponse
var resp clusterv1.NodeVMsResponse
err = json.Unmarshal(reply.Data, &resp)
require.NoError(t, err)
@@ -2300,7 +2300,7 @@ func TestHandleNodeVMs_UnknownInstanceType(t *testing.T) {
reply, err := daemon.natsConn.Request("spinifex.node.vms.unknown", nil, 5*time.Second)
require.NoError(t, err)
var resp types.NodeVMsResponse
var resp clusterv1.NodeVMsResponse
err = json.Unmarshal(reply.Data, &resp)
require.NoError(t, err)
spinifex/gateway/spx/spx.go
+4 -3
View File
@@ -9,6 +9,7 @@ import (
"strings"
"time"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/mulgadc/spinifex/spinifex/utils"
"github.com/nats-io/nats.go"
@@ -73,7 +74,7 @@ func GetNodes(ctx context.Context, nc *nats.Conn, expectedNodes int) (*GetNodesO
// VMInfoWithNode extends the daemon's VMInfo with node attribution.
type VMInfoWithNode struct {
types.VMInfo
clusterv1.VMInfo
Node string `json:"node"`
}
@@ -85,7 +86,7 @@ type GetVMsOutput struct {
// GetVMs queries all daemon nodes via NATS fan-out and returns their VMs.
func GetVMs(ctx context.Context, nc *nats.Conn, expectedNodes int) (*GetVMsOutput, error) {
frames, _, err := utils.Gather(ctx, nc, "spinifex.node.vms", []byte("{}"),
frames, _, err := utils.Gather(ctx, nc, clusterv1.NodeVMsSubject, []byte("{}"),
utils.GatherOpts{Timeout: 3 * time.Second, ExpectedNodes: expectedNodes})
if err != nil {
return nil, err
@@ -93,7 +94,7 @@ func GetVMs(ctx context.Context, nc *nats.Conn, expectedNodes int) (*GetVMsOutpu
allVMs := make([]VMInfoWithNode, 0)
for _, frame := range frames {
var nodeResp types.NodeVMsResponse
var nodeResp clusterv1.NodeVMsResponse
if json.Unmarshal(frame.Data, &nodeResp) != nil {
continue
}
spinifex/gateway/spx/spx_test.go
+7 -6
View File
@@ -4,6 +4,7 @@ import (
"encoding/json"
"testing"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/internal/testkit"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/nats-io/nats-server/v2/server"
@@ -119,11 +120,11 @@ func TestGetVMs_NoResponders(t *testing.T) {
func TestGetVMs_WithVMs(t *testing.T) {
_, nc := startEmbeddedNATS(t)
sub, err := nc.Subscribe("spinifex.node.vms", func(msg *nats.Msg) {
resp := types.NodeVMsResponse{
sub, err := nc.Subscribe(clusterv1.NodeVMsSubject, func(msg *nats.Msg) {
resp := clusterv1.NodeVMsResponse{
Node: "node1",
Host: "10.0.0.1",
VMs: []types.VMInfo{
VMs: []clusterv1.VMInfo{
{InstanceID: "i-abc123", Status: "running", InstanceType: "t3.small"},
{InstanceID: "i-def456", Status: "running", InstanceType: "t3.medium"},
},
@@ -149,10 +150,10 @@ func TestGetVMs_MultiNode(t *testing.T) {
for _, name := range []string{"node1", "node2"} {
nodeName := name
sub, err := nc.Subscribe("spinifex.node.vms", func(msg *nats.Msg) {
resp := types.NodeVMsResponse{
sub, err := nc.Subscribe(clusterv1.NodeVMsSubject, func(msg *nats.Msg) {
resp := clusterv1.NodeVMsResponse{
Node: nodeName,
VMs: []types.VMInfo{
VMs: []clusterv1.VMInfo{
{InstanceID: "i-" + nodeName, Status: "running"},
},
}
spinifex/handlers/eks/k3s_ha_control_plane.go
+4 -3
View File
@@ -11,9 +11,10 @@ import (
"github.com/aws/aws-sdk-go/aws"
"github.com/aws/aws-sdk-go/service/ec2"
clusterv1 "github.com/mulgadc/spinifex/contracts/cluster/v1"
"github.com/mulgadc/spinifex/spinifex/admin"
"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
"github.com/mulgadc/spinifex/spinifex/domains/ec2/instancetypes"
"github.com/mulgadc/spinifex/spinifex/foundation/aws/errors"
handlers_ec2_placementgroup "github.com/mulgadc/spinifex/spinifex/handlers/ec2/placementgroup"
"github.com/mulgadc/spinifex/spinifex/types"
"github.com/mulgadc/spinifex/spinifex/utils"
@@ -576,8 +577,8 @@ func (h *natsHostScheduler) InstanceHosts(ctx context.Context, instanceIDs []str
want[id] = true
}
out := make(map[string]string)
h.fanout(ctx, "spinifex.node.vms", func(data []byte) {
var resp types.NodeVMsResponse
h.fanout(ctx, clusterv1.NodeVMsSubject, func(data []byte) {
var resp clusterv1.NodeVMsResponse
if json.Unmarshal(data, &resp) != nil || resp.Node == "" {
return
}
spinifex/types/node.go
-36
View File
@@ -22,15 +22,6 @@ type GPUInfo struct {
Slices []GPUSliceInfo `json:"slices,omitempty"`
}
// VMGPUInfo describes one GPU attached to a VM.
type VMGPUInfo struct {
Model string `json:"model"`
VRAMMiB int64 `json:"vram_mib"`
PCIAddress string `json:"pci_address,omitempty"` // whole-GPU passthrough only
Profile string `json:"profile,omitempty"` // MIG profile name; empty for whole-GPU
MdevPath string `json:"mdev_path,omitempty"` // MIG only
}
// NodeStatusResponse is returned by the spinifex.node.status NATS topic (fan-out).
// Schedulable capacity = TotalVCPU - ReservedVCPU - AllocVCPU (same for memory).
type NodeStatusResponse struct {
@@ -76,30 +67,3 @@ type InstanceTypeCap struct {
MemoryGB float64 `json:"memory_gb"`
Available int `json:"available"`
}
// VMInfo describes a single VM for the cluster stats CLI.
type VMInfo struct {
InstanceID string `json:"instance_id"`
Status string `json:"status"`
InstanceType string `json:"instance_type"`
VCPU int `json:"vcpu"`
MemoryGB float64 `json:"memory_gb"`
LaunchTime int64 `json:"launch_time"`
// ManagedBy is the Spinifex platform component that owns this VM
// (e.g. "elbv2"). Empty for customer VMs. The UI uses this to filter
// system-managed resources out of customer-facing listings.
ManagedBy string `json:"managed_by,omitempty"`
GPUs []VMGPUInfo `json:"gpus,omitempty"`
// Health is a display label for instance health: "ok", "impaired",
// "recovering", or "-" for non-running VMs. CrashCount is the lifetime
// crash tally within the current restart window.
Health string `json:"health,omitempty"`
CrashCount int `json:"crash_count,omitempty"`
}
// NodeVMsResponse is returned by the spinifex.node.vms NATS topic (fan-out).
type NodeVMsResponse struct {
Node string `json:"node"`
Host string `json:"host"`
VMs []VMInfo `json:"vms"`
}